Heart failure and sudden death in dilated cardiomyopathy: a hidden competition we should not forget about when...

Area of Science:

• Cardiology
• Biostatistics
• Survival Analysis

Background:

• Dilated cardiomyopathy presents significant mortality risks from heart failure (HF) and sudden death (SD).
• Traditional survival analysis often assumes independence between competing events, which may not hold true in clinical settings.
• Accurate analysis of mortality in dilated cardiomyopathy requires methods that account for competing risks.

Purpose of the Study:

• To propose a protocol for analyzing competing risk events in dilated cardiomyopathy, mirroring univariate analysis approaches.
• To apply non-parametric bivariate survival estimators and a competing risk Cox model to mortality data.
• To investigate mortality patterns for heart failure and sudden death without assuming independence.

Main Methods:

• Utilized non-parametric bivariate survival estimators to model competing risks of HF and SD.
• Applied a multivariate proportional hazard model to a prospective cohort of 235 dilated cardiomyopathy patients (1978-1992).
• Stratified analysis by age, severity, and treatment, considering both intention-to-treat and actual treatment protocols.

Main Results:

• Bivariate survival curves revealed distinct mortality patterns for HF (early, then slowing) and SD (late increase).
• Competing risk analysis identified a treatment effect only under the actual treatment protocol.
• Standard Cox regression analysis did not reveal the same treatment effect, indicating potential misinterpretation.

Conclusions:

• Non-parametric bivariate survival estimation offers interpretable parameters similar to standard Cox regression.
• Accounting for competing risks is crucial to avoid misleading interpretations in dilated cardiomyopathy mortality studies.
• The proposed protocol provides a robust framework for analyzing competing risks in clinical research.